Model-Based Design for Affordability of a Netted Intelligence, Surveillance, and Reconnaissance Concept

advertisement
Model-Based Design for a Netted ISR Concept
Model-Based Design for Affordability of
a Netted Intelligence, Surveillance, and
Reconnaissance Concept
K. Dewayne Brown, Mary Beth A. Chipkevich, Robert J. Bamberger,
Tam-Thanh C. Huang, Mark A. Matties, James D. Reeves, and Christopher A. Rouff
ABSTRACT
This article highlights the critical importance of systems engineering methodology and its
influence on downstream outcomes in complex engineering concepts. It makes a development
case for complex systems of systems that contrasts netted with traditional intelligence,
surveillance, and reconnaissance. Model-based systems engineering methods supported by
development infrastructure can significantly impact the life-cycle affordability of complex
systems of systems. The long-standing systems engineering practices of the Johns Hopkins
University Applied Physics Laboratory (APL), the International Council on Systems Engineering,
and the Open Group Future Airborne Capability Environment Consortium, as well as the
objectives for projects such as the Defense Advanced Research Projects Agency Air Dominance
Initiative, provide context for an APL brand of model-based methods. This article introduces
an APL model-based systems engineering methodology within an integrated development
environment and discusses the methodology in the context of a netted intelligence, surveillance,
and reconnaissance concept.
INTRODUCTION
BACKGROUND
More than 70% of program life-cycle costs are determined in the concept engineering phases. A modelbased systems engineering (MBSE) methodology,
integrated development environment (IDE), and systems
engineering (SE) infrastructure provide engineers the
capability to cost-effectively explore, model, prototype,
and validate concepts. A case study for intelligence, surveillance, and reconnaissance (ISR) concept engineering piloted use of this capability, and it was determined
that one concept was more affordable (in terms of lifecycle costs) and rapidly realizable (in terms of development time lines) than an alternative concept.
The overall complexity of national security and operational warfare is increasing.1 Complexity of the systems
of systems (SoS) needed to deliver mission capability is
also increasing, particularly with regard to the amount of
information exchanged, the degree of interaction among
battle force units, and the battlespace environment.2
Concurrently, DoD budgets are trending downward.1
Affordability is the new number one threat to developing, delivering, integrating, and sustaining needed capabilities.3 The DoD has recognized the need to reduce
life-cycle costs early in the development of concepts and
capabilities, as evidenced by various studies and initiatives (see Fig. 1).
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
23­­­­
K. D. Brown et al.
The Reality...
“Our current security challenges are more formidable and
complex than those we faced in downturns following Korea,
Vietnam, and the Cold War. There is no foreseeable ‘peace
dividend’ on our horizon.”
—GEN DEMPSEY, CJCS
Testimony to SASC, 12 Feb 2013
$800
Korea
Vietnam
OEF/OIF
Reagan Buildup/
Cold War
$700
$500
$400
1996
2020
2017
2014
2011
2008
2005
2002
1999
1996
1993
1990
1987
65-year Average
= $477B
1984
1978
1975
1972
1969
1966
1957
1954
1948
$0
1951
$100
1963
Actual
65-year Avg.
PB13 Projection
Sequestration Projection
Historical Trough Projection
$200
1981
$300
1960
CY13 $ in B
$600
Figure 1. Battlespace complexity and affordability. [Right image from G. Boltz (APL) and left from Ref. 1.]
MOTIVATION
Applying SE practices provides necessary structure
throughout the life cycle of a system. Supporting modelbased infrastructure enables effective management of
performance, schedule, and cost over significant portions
of the concept-development life cycle, including exploration, verification, and validation. Models facilitate
efficient expression of system requirements, structure,
behavior, and function along with rapid communication,
distribution, and functional decomposition of complex
systems. Model-based methods enable effective, rapid,
and low-cost decisions to be made among alternatives
when engineering a complex SoS. In this case study,
application of a model-based methodology demonstrates
the ability of these methods to significantly impact the
life-cycle affordability of ISR concepts.
Figure 2. A maritime interdiction scenario.
24­­­­
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Model-Based Design for a Netted ISR Concept
Figure 3. Anti-access/area-denial maritime dynamic targeting scenario.
THE ISR CASE: THE CHALLENGE
Warfighters need a capability to command and operate an ISR SoS more effectively and affordably than they
can today. Tasking–collection–processing–exploitation–
dissemination (TCPED) is a process that uses an SoS
during military operations to achieve ISR mission objectives determined by commander’s intent. The tasking–
collection (TC) portion of TCPED encompasses actions
and decisions related to planning for, tasking of, and
operation of systems that contribute to ISR mission
objectives. The increasing number and diversity of ISR
assets as well as competing ISR mission requirements
challenge the effectiveness of existing doctrine, processes, and systems to command and manage the use of
ISR platforms and sensors.4, 5 The volume and multiple
modalities of sensor data can overwhelm the processing–
exploitation–dissemination (PED) portion of TCPED.5
“As-is” traditional ISR SoS TCPED processes have
time-consuming and manpower-intensive TC and PED
activities and require high-bandwidth reachback communications. As-is TC requires mission control operators at ground control stations or on manned airborne,
surface, or subsurface platforms to perform sensor control and platform control of assigned “stovepiped” ISR
assets in a reactive rather than a predictive manner.
This reduces time available to properly respond to a
situation and suboptimizes overall ISR SoS performance. As-is PED involves transfer of raw sensor data
via reachback communications to intelligence community facilities. Analysts use available systems at these
locations to process, exploit, and disseminate fused
intelligence products for use by command-and-control
decision makers.
In dense battlespace environments and with increasingly complex SoSs, synchronization among nodes in
the SoSs operated by naval, joint, and coalition forces
requires timely and tight coupling with operations, targeting, and intelligence operational doctrine to deliver
the ISR targeting information necessary to support weapons planning.5, 6 This is particularly true for dynamic
targeting scenarios. As-is TCPED with traditional ISR
SoSs does not provide the required timely coordination
and reporting between operational processes and nodes/
capabilities.2, 4, 5
Figures 2 and 3 show notional scenarios for maritime interdiction and anti-access/area-denial maritime
dynamic targeting, which provide operational context
with communications, threat detection tracking, classification, and other challenges for the netted ISR concept.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
25­­­­
K. D. Brown et al.
THE ISR CASE: NETTED ISR CONCEPT
The concept of a “to-be” TCPED-centric netted ISR
SoS is an alternative to the present as-is ISR architecture. A TCPED-centric netted ISR SoS enables the
warfighter to command a collaborative sensor fusion
capability and optimally, reliably, and locally control
actions and decisions for each node; it also increases the
ability of the SoS accomplish missions and tasks. Data
fusion (e.g., PED), previously completed in central locations, now occurs on sensor platforms. It uses organic
and remote sensor data, increasing resiliency in degraded
communications environments. Data fusion also occurs
at data fusion centers specified by the commander.
Constellation-wide sensor resource allocation (e.g., TC)
occurs on both platforms and among operational commands designated by the commander to optimize data
collection. The output of data fusion is fed back nearly
instantaneously as input to sensor fusion. The netted
ISR concept is enabled by end-to-end connectivity with
quality of service that allows in-degree nodes of the SoS
to assuredly share relevant information.
With netted ISR, command, sensor fusion, and data
fusion capability is allocated to functional elements on
segments that distribute across physical nodes in the
SoS. Segments are a collection of functional elements.
The expected role of each node determines the extent
to which each segment is instantiated on the node. The
designated role of each node for a specific operational
mission determines the extent to which the instantiated
segments are used on each node. As such, the netted ISR
concept is extensible to platforms with varying space,
weight, and power constraints; scalable to a very large
number of connected nodes; and adaptable to a range of
collaborative behaviors (e.g., nodes behaving as individual contributors, as neighbors, as a community, or with
a swarm identity).
Functional elements of command include the ability to collaborate, translate the commander’s intent
into measures allocated to tasks and objectives, assess
the performance of the SoS and infer achievement of
the commander’s intent, decide the priority of mission
requirements to achieve the commander’s intent, and
assign mission requirements to ISR assets.
Functional elements of sensor fusion include the
ability to infer sensor collection needs, decide and task
sensor data collections and platform movement, skew
sensors, move sensor platforms, collect sensor data, and
share sensor data.
Functional elements of data fusion include the ability
to receive and prepare sensor data for processing; screen
sensor data; share relevant sensor data; perform processing
to detect objects, track hostiles, precisely track hostiles,
track all objects, classify targets, and determine intent of
targets; develop object/target reports; share object/target
reports; and maintain an ISR data repository.
26­­­­
Implementation of data fusion includes data conditioners, screening components specialized to multiintelligence and multi-modality sensors, fusion
components that perform data association, kinematics
state estimation, class estimation, and data association.
The data fusion approach exploits complementary attributes of diverse sensor phenomenology/geometries and
data collected over time to maintain a low system-level,
post-correlation false-alarm rate. Output conditioning
prepares and disseminates fusion output (e.g., actionable information, target reports, and target nomination
reports) at achievable data transmission rates.7
Implementing sensor fusion will include automated
components that coordinate and synchronize ISR sensor
platforms as an integrated unit to continuously maximize aggregate net fused information gain and adjudicate tasking among competing priorities across the
entire search volume and all targets, as well as over a
configurable finite planning time horizon.8
Implementing command of a netted ISR SoS will
involve automation that shifts much knowledge and
inference functionality performed by operators in the
cognitive domain to processors in the logical domain.
Collaborative cognition will occur between warfighters
and the command capability of the netted ISR SoS. The
netted ISR command functional element will provide
the capability to reliably optimize responses to emergent
behaviors associated with SoS complexity;9 relate those
responses to commander’s intent, battlespace complexity, cognitive parameters, and decision policies established in doctrine; and synchronize the information,
social, cognitive, and physical domains of battlespace
management command and control.10
The concept of netted ISR and its collaborative
command, sensor fusion, and data fusion capability
aligns well with the Navy Information Dominance Roadmap, which states, “Battlespace Awareness will require
enhanced information content, advanced means to rapidly sense, collect, process, analyze, evaluate and exploit
intelligence regarding our adversaries and the operating
environment.”11
Netted ISR Reference Architecture
To facilitate SE of the netted ISR concept, a reference architecture was developed, as depicted in Fig. 4.
The reference architecture is intended to guide and
frame architecture and solution development. There
are sensor, command, and fusion segments, which are
instantiated on nodes that distribute across an SoS and
interconnect by a heterogeneous network.
All node types have a multiprocessing hardware architecture. The processors can be subdivided into non-realtime and near-real-time types. Software that is event
driven, such as user/mission applications and higher-level
networking services (application, presentation, session,
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Model-Based Design for a Netted ISR Concept
Sensor
subsystem
Sensor 1
Sensor N
Sensor
manager
Platform
manager
Platform
control
subsystem
Local sensor data fusion
applications
Command applications
Local
sensor
fusion
Local core services
Network services
Communications services
Task
sensors
Task
weapons
Task
platforms
Decide
Infer
Translate
intent
Network services
Communications services
Local core services
Hardware
Hardware
Sense segment ∑im (m = 1, 2,..., M)
Heterogeneous
network
Command segment ∑in (n = 1, 2,..., N)
ISR fusion applications
Sensor
data
fusion
ISR
repository
Local core services
Sensor
fusion
Common
operating
picture
Network services
Communications services
Hardware
Fusion segment ∑ip (p = 1, 2,..., P)
Figure 4. Netted ISR reference architecture.
transport, and network OSI [Open System Interconnection] layers), are hosted on the non-real-time processors,
such as general-purpose processors and/or graphic processor units. The software architecture can be partitioned
into two broad groups. Software that is near real time,
such as lower-layer networking services (data and physical OSI layers), is hosted on near-real-time processors,
such as digital signal processors or field-programmable
gate arrays. The software architecture can be partitioned
into user/mission applications and services such as local
core, networking, and communication. The local core,
networking, and communication services enable applications to execute the mission and interact with any node
distributed in the SoS and across all segments of the
SoS. User/mission applications are specific to each type
of segment. For instance, sensor segments host sensor
interfaces, whereas the command segment hosts user
interfaces and fusion segments host data-repository interfaces. This reference architecture is modular and scalable,
which facilitates rapid integration of innovative technology, legacy design reuse, and interoperability between
diverse functional platforms and heterogeneous networks.
Case Study Approach: Netted ISR Model-Based Design
for Affordability
With the ISR challenge defined, operational context
provided, and initial netted ISR concept and reference
architecture developed, the SE challenge is to validate
that the netted ISR concept can realize the needed ISR
capability more affordably and rapidly than the traditional ISR of today.
Resource constraints on the MBSE methodology
development were a 4-month schedule and eight parttime subject-matter experts. The study team focused
on design for affordability of the netted ISR concept
and performed one innovation cycle using the MBSE
methodology and SE infrastructure. The approach
included modeling and analysis of ISR performance
for an operationally relevant scenario as well as rapid
development of an ISR network model, which enabled
efficient exploration of both the traditional and netted
ISR concepts. The network model was central to rapidly
establishing ISR prototypes of both the traditional
and netted concepts. Preliminary verification tests
demonstrated performance of the ISR concepts. A cost
model was developed and used to assess the affordability
of both concepts.
Results of Netted ISR Case Study
Results from one innovation cycle of the concept
engineering phase showed that the netted ISR concept
directly improved the affordability and effectiveness of
ISR. Preliminary verification tests for one scenario demonstrated that the netted ISR concept required 30%
fewer sensor platforms and at least 25% fewer operators than the traditional ISR model. Netted ISR can
reduce operational costs and reduce the number of assets
needed to perform ISR. Cost modeling and affordability
analysis showed significant, compounded life-cycle savings with the netted ISR concept compared to the traditional concept, especially as the battlespace becomes
more complex.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
27­­­­
K. D. Brown et al.
Mission/operational
architecture and capabilities
Solution (systems and services)
architecture and functionality
Solution
validation
Integrated System Model
ta
duc
it
rch
ect
ign
des
nd
a
ure
This case study made use of an MBSE methodology,
IDE, and SE infrastructure to design for affordability as
part of a 4-month concept engineering phase of a complex SoS. The remainder of this article describes the
Johns Hopkins University Applied Physics Laboratory
(APL) MBSE methodology, IDE, and SE infrastructure
used for this case study and its relevance to supporting
and influencing SE over the life cycle of a system.
on
Pro
Figure 5. APL SE process. (Reprinted from Ref. 12.)
Detailed
design
Integration,
test, and
verification
Implementation
Figure 6. INCOSE MBSE methodology. HW/SW, hardware/software. (Reprinted from Ref. 13.)
APL SE Process
Implementation of the netted ISR architecture
requires an effective balance of scientific and engineering principles, performance, requirements, and sponsor
program constraints. This is a classical SE challenge.
The APL SE loop,12 which has matured over decades,
is shown in Fig. 5, and the major phases used to solve
national and international critical SE challenges are
described in Table 1. Every step produces knowledge
and experience, which is fed back into subsequent spirals. The APL SE process is mature and proven and has
informed the development of the APL MBSE IDE.
Table 1. APL SE Process
28­­­­
Mission/operational
Operation and
architecture views
maintenance
Systems and services
architecture views
Systems
engineering
Requirements views
System
and
HW/SW verification and
views
architecture
validation
Concept of
operations
tion
xplora
Concept e
gra
ti
Systems Engineering
Capability
integration
inte
Critical
needs
Capability
assessment
Capability
architecture
Solution
implementation
and
nt
Pro
duc
t te
st
yme
Deplo
Process Step
Description
Critical needs
Operational/mission data analysis is conducted, focused on establishing concept feasibility and exposing critical needs.
Capability
assessment
Existing systems are evaluated to determine
whether they can meet the need or whether a
new capability development is required.
Concept
exploration
Alternative candidate concept designs, models,
and analyses are completed. Trade analysis
enables effective down-selection to a best concept based on any number of metrics, such as
performance, efficiency, economy, risk, utility,
or a combination of these characteristics.
Validation
Proof-of-concept (PoC) prototypes are developed to verify functional performance in a
representative environment.
Implementation
The concept is realized in an operational prototype, and verification tests are completed.
Deployment
The concept is deployed for field test
validation.
INCOSE MBSE
MBSE provides a methodology to more effectively
manage the SE challenge at lower cost and shorter
development cycles. The International Council on Systems Engineering (INCOSE) has developed an MBSE
methodology,13 which is shown in Fig. 6. This diagram
highlights how models are central to this methodology. The Integrated System Model is a repository for
all knowledge about the system (aiding in communication and collaboration). Capability and product architectures, along with design, verification, and validation
testing, are derived from the Integrated System Model.
This results in consistency, which improves maintainability and reduces ambiguity. Requirements traceability
Government
partners
IDE
MBSE methodology
Development Framework
Development
framework
Development
tool set
Project
repository
Reference
architecture
Commercial
partners
Figure 7. Integrated development environment.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Model-Based Design for a Netted ISR Concept
Sponsor challenge
System
PoC
prototyping
MBSE
IDE
Reference architecture
Concept exploration: M&S
Validated concept
HW/SW prototyping
System concept
verification
System concept
implementation
Development tools
System concept
validation
IDE framework
System M&S
concept
exploration
IDE
backplane
V&V tools
Figure 8. APL MBSE methodology. M&S, modeling and
simulation.
to architecture and associated design elements enables
change impact analysis. Documentation is generated
from the Integrated System Model, ensuring accuracy.
Automatic code-generation processes ensure efficient
implementation and quality management. Because the
INCOSE MBSE developments are embraced by a broad
range of SE organizations, the INCOSE model has
informed the APL MBSE IDE development.
MBSE IDE
APL is maturing an MBSE IDE that enables distributed stakeholders to collaboratively contribute solutions
to the SE challenge. The APL SE IDE is initially provisioned and targeted for the concept exploration, implementation, validation, and deployment SE processes. It is
relevant for application by projects such as the Defense
Advanced Research Projects Agency-sponsored Air
Dominance Initiative project,14 whose objective was to
integrate its Strategic Technologies Office capabilities for
communications in a contested environment. The goals
of the Air Dominance Initiative project included integration of legacy and future airborne communication waveProvides a
rich system
design and
analysis
environment
Design flows
provide
machinegenerated
code
Concept system Machinemodeling and generated
simulation
code
(blade servers)
Test results
feedback to
system concept
Provides a
scalable collection
of reconfigurable
multiprocessor
architectures
Reconfigurable Prototype
processing hard- models
ware prototypes to industry
(SDR 4000)
V&V testing/field
demonstrations
(programmable
test equipment)
Industry product
implementation
Provides a rich
collection of
programmable test
vectors and
measurements
Repository
Figure 10. APL IDE framework. HW/SW, hardware/software.
forms through heterogeneous networking protocols across
increasing numbers of pervasive airborne nodes while
achieving drastically reduced innovation cycles. The APL
IDE as shown in Fig. 7 facilitates distributed collaboration on shared centralized resources among government
sponsors, users, and technical leadership with commercial hardware, software, and service providers. It is based
on an integration of MBSE methodology, a distributed
integration framework, hardware and software reference
architectures, development toolsets, and a repository.
AN APL-Branded MBSE Methodology
An APL-branded MBSE methodology (APL has
developed a unique collection of processes, methods, and
tools) is shown in Figs. 8 and 9. The major iterative and
collaborative processes include (i) concept exploration
through modeling and simulation, (ii) rapid prototyping
through code generation, (iii) verification and validation
(V&V) testing, and (iv) submission to the project repository. An inner closed loop enables rapid verification of
concepts developed in the low-cost-of-change (relative
Reference
HW/SW
architectures/
design kits
Concept
exploration
tools
Application,
network, and
communication
services development tools
Process capabilities feedback to system concept
Figure 9. APL MBSE methodology flow.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Concept
V&V tools
Repository
tools
IDE framework
Figure 11. APL IDE notional user interface.
29­­­­
K. D. Brown et al.
to cost of change in physical hardware and software)
modeling and simulation
environment and influenced by lessons learned
from V&V testing. The
outer closed loop enables
stakeholders to influence
the concept development
through disciplined design
for X (where X is a variable
for affordability, six-sigma,
testing, manufacturing, and
so forth).
Netted ISR master reference architecture
Software computing architecture
Portable components segment (applications)
VM
Sensing applications
Fusion applications
Sensor 1 Sensor 2 Sensor N
Process 1 Process 2 Process N
Platform-specific services segment
(local core services)
Platform common services
Config
Logging
Platform device services
App 1
App N
App 2
Transport services segment
(network/communications services)
VM
Distribution (Dist) capabilities
WF
functions
Dist
Virtual
machine
(VM)
GPS
Data-processing applications
WF fcn
transport
QoS
Interworking fcn
Config
Secured
Red
data service routing
Red
QoS
IDE Framework
The SE IDE framework
shown in Fig. 10 illustrates
how hardware and software reference architectures, development, V&V
tools, and a repository can
be integrated. This integration facilitates transfer of
information,
availability
of resources, and maturation of concepts. Shown in
Fig. 11 is a notional framework interface that enables
distributed developers to
navigate, access, and process
using shared IDE resources
for concept exploration,
implementation, verification, and validation. Within
the IDE, integrated tools are
able to interoperate across a
common backplane. Future
tool versions will be provisioned with plug-and-play
interfaces, common reference architecture templates,
and repository libraries.
VM
Input/output
services segment
(networking services)
Addressing
services
Power
control
Mobility
management
Black
routing
Neighbor
discovery
Black QoS
Flow
control
Operating system (OS) segment
OS kernel
Partitioning
File system
Hypervisor
Hardware architecture
Digital hardware
Waveform processor
Device
driver
Device
driver
Application/
networking processor
Switch fabric
Backplane switching
Analog hardware
Link-16
TTNT
MADL
CDL
Future
Figure 12. APL master reference architecture. CDL, common data link; Config, configuration; DNS,
domain name service; fcn, function; MADL, multifunction advanced data link; QoS, quality of service; WF, web file.
Hardware and Software Reference Architecture
Within the SE IDE is a hardware and software master
reference architecture, as shown in Fig. 12; it is based
on the Technical Standard for Future Airborne Capability Environment (FACE)15 and provides functional and
interface definitions for the hardware and software to be
implemented. It illustrates the major software application, networking, and communication services as well
as the major non-real-time and near-real-time hardware
processor types. Incremental implementation of this ref-
30­­­­
DNS/subnet
formation
erence architecture is planned for several APL projects
over the next few years.
Concept Exploration
The SE IDE includes modeling and simulation toolsets
that enable concept exploration through PoC modeling
and simulation. These concept virtual prototypes (VPs)
enable exploration of a range of alternative approaches,
analyzing the performance, costs, and benefits of each
approach in an agile and low-cost-of-change environment.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Model-Based Design for a Netted ISR Concept
Pulse n: 256 encoded bits + 24 bits (S1) +
24 bits (S2) + 105 bits (NTmin) +
max 512 bits (corresponding to largest Ln) =
(max) 921 bits into GMSK
(10×1)
Bernoulli
binary
(10×1)D2
(10×1)
Select Packet Size = [size, delay]
using Callback InitFcn
in model properties
Packet
Cyclic
encoder
(5×1)D2
(549×1)D3
Carrier freq
Assemble
various seed packets
at mbps
FH-CPM
modulator
Frequency
hopping GFSK
modulator
(54900×1)D3
(54900×1) D3
(54900×1)
Upconverter
AWGN
Carrier freq
nfmt19937af,
“seed”
P library
(54900×1)
(54900×1)
Downconverter
D
(54900×1)D
To
frame
(54900×1)
(549×1) Disassemble (5×1)D2
FH-FM
Cyclic
packet
demodulator
decoder
Frequency hopping
Shortened
FM demodulator
hamming (15×10)
decoder
Receiver
D
15-FSK
(54900×1) D
Carrier freq
Generate
(54900×1)
15 possible carriers
–12 MHz to 12 MHz
D
6
U
Hop frequency in MHz
Frequency hopping code
Channel
Channel
Transmitter
(549)
Channel
(54900×1)
(5×1)D2
RxPacket
Open
scopes
Spectrum Close
scope scopes
Received signal
spectrum
(10×1)
0
s1 del (10×1) D2
Tx
Align s2 (10×1) D2
0
Error rate 3 D2
s2 signals
D5
RxPacket
calculation
5.253e+05
delay
Rx
(10×1)
Carrier
BER
(10×1)
Align signals
frequency
BER calculation
monitor
380
Delay
Hop time will vary by changing packet size
TxPacket
s1
AWGN, additive white Gaussian noise; BER, bit error rate; FH-CPM, frequency hopping
continuous phase modulation; FH-FM, frequency hopping frequency modulation; freq,
frequency; FSK, frequency shift keying; GFSK, Gaussian frequency shift keying; GMSK,
Gaussian minimum shift keying; Ln, length; NTmin, number of minimum time periods;
PN, pseudo random; RxPacket, receive packet; TxPacket, transmit packet
Figure 13. TTNT Simulink physical layer modulation virtual prototype.
Figure 14. Netted ISR maritime interdiction scenario virtual prototype.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
31­­­­
K. D. Brown et al.
C-IDE
MATLAB
system
model
Real-time
workshop/
embedded
coder
Simulink
system
model
System
generator
DSP
compiler/
linker
DSP
device
programming
GPP
compiler/
linker
GPP
device
programming
VHDL
compiler/
linker
VHDL
device
programming
Reconfigurable
multiprocessor
prototype
Executable build
HDL IDE
Executable code
Concept M&S analytical TB
Source code
As a second example,
consider the netted ISR
communication network SM
for the maritime interdiction
scenario shown in Fig. 14.
This VP enables exploration
of operational performance
sensitivity to communication
link parameters such as quality of service, throughput,
range, power, and bandwidth.
Each of these exemplar virtual prototypes enabled rapid
exploration of critical performance parameters in a lowcost-of-change environment.
Concept Validation
The SE IDE also includes
source-code-generation toolsets for rapid prototyping
(i.e., converting system SMs
developed during concept
exploration into a PoC physical prototype [PhP] whose
response to physical environmental processes can be
measured with calibrated equipment). The calibration
enables quantification of SM confidence intervals when
measurements are compared to simulated responses. To
enable rapid, low-cost-of-change prototypes, rapid software prototype tools are used, as shown in Fig. 15. These
tools convert SM into source code, which can be compiled into executable code for target processors.
Figure 15. Rapid prototype auto-generation processing. DSP, digital signal processor;
GPP, general-purpose processor; HDL, Hardware Description Language; M&S, modeling and simulation; TB, testbed; VHDL, Very High-Speed Integrated Circuit HDL.
Alternative down-selection and optimization highlight
how rapid concept engineering is achieved.
As an example, consider a Tactical Targeting Network Technology (TTNT)16 surrogate modulator
model, as shown in Fig. 13. This simulation model (SM)
is only a portion of the communication physical layer
processing but highlights the capability to test system
performance sensitivity under a range of modulation
parametric variations.
NISR master reference architecture
Requirements
Software computing architecture
Portable components segment (applications)
Sensing applications
Platform common services
Logging
Config
Platform device services
VM
Data-processing applications
Process 1 Process 2 Process N
App N
App 2
Reference
architecture
VM
Distribution capabilities
Dist
WF fcn
transport
DNS/subnet
formation
Addressing
services
App 1
Transport services segment
(network/communications services)
WF
functions
Virtual
machine
(VM)
GPS
I/O services segment
(networking services)
VM
Fusion applications
Sensor 1 Sensor 2 Sensor N
Platform-specific services segment
(local core services)
QoS
Interworking fcn
Power
control
Config
Secured
Red
data service routing
Mobility
management
Black
routing
Red
QoS
Constraints
Black QoS
Neighbor
discovery
Laboratory or demonstration
environmental testbed
Flow
control
Operating system segment
OS kernel
Partitioning
File system
Device
driver
Hypervisor
Hardware architecture
Digital hardware
Waveform processor
Device
driver
SysML updates/
modifications
Application/
networking processor
Switch Fabric
Backplane switching
Analog hardware
Link-16
TTNT
MADL
CDL
Calibrated
V&V
testing
Future
Automated
code
generation
Reverse
engineering
SysML tools
(MagicDraw,
Rhaposody, etc.)
Simulationdriven
development
1
s
Vc
RC
+
3
Ro
+
÷
×
×
1
S
S
_
+
×
+_
mode
4
Io
>= 0
OR
== 1
1/C
Vc
×
÷
+
+
RC
Vo
1
y
2
Vs
== 2
OR
–
+
–
double
vL
1
s
1/L
IL
IL
<0
Is
×
0
Test
lessons
learned
Terminal TX
system model
validation
Difference
Concept
exploration
Uplink
channel
impairments
Terminal
TX
Model
response
System model testbed
Ic
double
Hardware
response
Terminal TX
hardware
in the loop
Executable
code
Custom
code New source
tracking
code
Legacy
source code
Concept validation
Base station
RX hardware
Downlink
channel
impairments
Satellite
RX
Satellite
processing
Satellite
TX
RL
Simulink
Figure 16. PhPs by automatic code generation.
32­­­­
Figure 17. Hardware/software-in-the-loop
RX, receiver; TX, transmitter.
(SiL)
V&V
TB.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Model-Based Design for a Netted ISR Concept
Stimulus
equipment
Measurement
equipment
Wireless
MIMO channel
emulator
RF
signal
generator
Digital
oscilloscope
Large-/small-scale
fading
Data
pattern
generator
TX
UUT
Function/
noise
generator
Vector
signal
analyzer
RX
UUT
Multipath
fading
Digital
logic
analyzer
Doppler
spread
BERT
Network
analyzer
Noise/
interference
Figure 18. Communication V&V TB. BERT, bit error rate tester; MIMO, multiple-input and multiple-output; RX UUT, receive unit under
test; TX UUT, transmit unit under test.
SIL
(10×1)
Bernoulli
binary
(10×1)D2
TxPacket
Cyclic
encoder
D3
FH-GMSK Demod
(5×1)D2
(549×1)D3
AWGN
Channel
Transmitter
Channel
(54900×1)D3
Carrier freq
(54900×1)D4
(54900×1)D3
(54900×1)
Carrier Freq
Assemble
various seed packets
at 1 Mbps
(54900×1)
(54900×1)
Channel
Open
scopes
Spectrum Close
(5×1)D2
Disassemble (5×1)D2
scope scopes
Cyclic
RxPacket
packet
decoder
Received signal
Shortened
FH-GMSK Demod
spectrum
hamming (15×10)
decoder
Receiver
0
D3
(10×1)
s1 del (10×1) D2
Tx
SIL
TxPacket
s1 Align
3 D2
(10×1) D2
0
Error
rate
s2
(10×1) s2 signals
RxPacket
calculation
D5
5.13e+04
delay
FH-GMSK Demod
Rx
BER
(10×1)
Align signals
BER calculation
(54900×1) D4
750
D4
D4
nfmt19937af,
To
Carrier freq
Delay
15-FSK
“Seed”
frame
Seed
Generate
(54900×1)
15 possible carriers
–12 MHz to 12 MHz
D4
6
U
Hop frequency in MHz
Carrier
frequency
monitor
AWGN, additive white Gaussian noise; BER, bit error
rate; freq, frequency; FSK, frequency shift keying;
GMSK, Gaussian minimum shift keying; RxPacket,
receive packet; TxPacket, transmit packet
Frequency hopping code
Figure 19. TTNT physical layer modulation SiL.
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
33­­­­
K. D. Brown et al.
Reference Architecture SysML
Rapid Prototyping
An MBSE rapid software prototyping example includes a processing hardware and software reference
architecture such as FACE, performance requirements, and resource
constraints. System designers can
leverage legacy code to prototype
new source code. As shown in
Fig. 16, the SE IDE provisioned with
SysML (Systems Modeling Language) and UML (Unified Modeling Language) toolsets (reference
templates and libraries) is capable
of leveraging legacy design, integrating new requirements within a
Figure 20. TTNT physical layer modulation PiL.
reference architecture, and generating new source code. This example
demonstrates how legacy code is
Fig. 18 can be integrated with the concept exploration
reverse-engineered by a SysML17 tool, such as MagicDraw
environment by applying the same stimulus used in the
(http://www.nomagic.com/products/magicdraw.htm) or
SM, and responses from the V&V TB can be compared
Rhapsody (http://www-142.ibm.com/software/products/
to the simulated responses in the SM.
us/en/ratirhapfami/). The reverse engineering produces
Consider the rapid prototyping example in Fig. 19;
SysML diagrams of what the code represents and how
it is a software-in-the-loop (SiL) emulation, which is
they are interconnected and can be reused in the new
derived from the previous TTNT modulation SM. The
system design. Based on the requirements and reference
SiL provides a capability to emulate the behavior of
architecture, the system designer develops a model of the
18
source code generated from the SM.
system in SysML and UML. Throughout the system
As a second example, consider the processor-in-thedesign process, Simulink (http://www.mathworks.com)
loop (PiL) shown in Fig. 20; it is a digital signal processor
code can be generated/developed to simulate target comhardware emulation derived from the same SM. It proponents. After the system has been modeled, code that
represents the system is automatically generated. Custom
vides a capability to embed physical hardware running
code can be written for any components where automatic
the SiL software within the SM. Implementation errors
code generation is not possible. The SysML tool then
associated with fixed-point precision, timing, and clocktracks any custom code inserted or added to include in
ing can be evaluated with the PiL. Both the SiL and
future automatic code generation.
Automatic test cases are also generLaptop-UAV
ated from the requirements stored
in SysML. The automatic code
generation can also be done for a
partial system model or for subsystems to test the system as the model
is being developed to support an
incremental or spiral development.
The PoC PhP can be stressed
SBC-Command
by mechanical, electrical, or
environmental conditions in an
integrated hardware/software-inthe-loop V&V testbed (TB), as
shown in Fig. 17. These tests can
SBC-Sensor
be conducted within a laboratory
Laptop-Cutter
Laptop-UAV
or extended to include field-test
demonstrations. For instance, the
communication V&V TB shown in
Figure 21. Netted ISR physical layer prototype. SBC, single-board computer.
34­­­­
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Model-Based Design for a Netted ISR Concept
MBSE capability
Reduced cycle times
System of systems
interoperability
Design optimization across broad trade space
Cross domain effects-based analysis
Extending maturity and capability
Institutionalized
MBSE across
academia/
industry
Defined MBSE theory, ontology, and formalisms
Maturity
Well-defined
MBSE
Distributed and secure model repositories
crossing multiple domains
Matured MBSE methods and metrics,
integrated system/HW/SW models
Ad hoc MBSE
document
centric
Refer to activities in
the following areas:
Architecture model integrated with
simulation, analysis, and visualization
Emerging MBSE standards
2010
• Planning and support
• Research
• Standards development
• Processes, practices, and methods
• Tools and technology enhancements
• Outreach, training, and education
2020
2025
Figure 22. INCOSE MBSE methodology maturity. (Reprinted with permission from Ref. 19, © INCOSE.)
the PiL were auto-generated and compiled in minutes.
Changes to the SM can be rapidly transferred to both
emulations for verification testing.
As a third example, consider the netted ISR network
PoC PhP shown in Fig. 21. In this case, physical hardware
has been connected to nodes in the previously discussed
VP. Operational hardware hosting operational software
can execute in real time. Communication traffic is generated by the PhP, exchanged between the VP nodes, and
processed by the destination PhP nodes. In this manner,
the hardware-in-the-loop enables verification of node
application, networking, and communication services
while retaining a low-cost-of-change environment for
the location, ranges, and wired and wireless link effects.
Link latency, quality of service, and throughput can rapidly and easily be manipulated to quantify performance
sensitivities of an operational scenario.
Integrated Validation Environment
The SE IDE PhP can be extended to field demonstrations (for example, consider the UAV). The netted ISR
PoC PhP can be deployed on a constellation of airborne
platforms and demonstrated under actual operational
conditions to validate the concept. A similar test stimulus used in the SM can be applied to the field demonstration hardware, and responses can be measured and
compared to the simulated responses in the SM.
CONCLUSION
The APL MBSE IDE has been introduced and
applied to the anti-access/area-denial netted ISR case
and a growing list of other SE projects. The IDE infra-
structure facilitates effective, agile, rapid, and affordable concept exploration, verification, and validation
for complex SoS such as netted ISR. For the netted ISR
case, a rapid innovation cycle project over the course
of 4 months used an IDE to define, model, prototype,
and produce initial results that indicated that 30% fewer
sensor platforms and at least 25% fewer operators were
necessary for one scenario, contributing to the potential for significant, compounded life-cycle savings with
netted compared to traditional ISR SoS.
Studies conducted by the National Defense Industrial Association, INCOSE, and other SE organizations
recommend adoption of MBSE methodologies based on
member-documented project cost and schedule reductions.19–21 The INCOSE MBSE roadmap is shown in
Fig. 22. APL is midway through maturing its brand of
MBSE methodology.
REFERENCES
1Shaffer,
A., Current S&T Priorities and the Future of DoD S&T,
Department of Defense, Washington, DC (29 Oct 2013), http://
www.acq.osd.mil/chieftechnologist/publications/docs/Current_ST_
priorities_and_the_Future_of_DOD_ST.pdf.
2Office of the Under Secretary of Defense for Acquisition, Technology, and Logistics, Report of the Joint Defense Science Board Intelligence
Science Board Task Force on Integrating Sensor-Collected Intelligence,
Defense Science Board, Washington, DC (Nov 2008).
3Dunaway, D., “Creating Integrated Warfighting Capabilities,”
Proceedings Magazine 139/8/1,326 (2103).
4Joint Operational Access Concept (JOAC), version 1.0, Joint Chiefs of
Staff, Washington, DC (Jan 2012).
5United States Joint Forces Command, Commander’s Handbook for
Persistent Surveillance, version 1.0, Joint Warfighting Center, Joint
Doctrine Support Division, Suffolk, VA (20 Jun 2011).
6Joint Publication 3-60 (JP 3-60), Joint Targeting, Joint Chiefs of Staff,
Washington, DC (31 Jan 2013).
7Newman, A. J., and Mitzel, G. E., “Upstream Data Fusion: History,
Technical Overview, and Applications to Critical Challenges,” Johns
Hopkins APL Tech. Dig. 31(3), 215–233 (2013).
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
35­­­­
K. D. Brown et al.
8Newman,
A. J., and DeSena, J., “Closed-Loop Collaborative Intelligence, Surveillance, and Reconnaissance Resource Management,”
Johns Hopkins APL Tech. Dig. 31(3), 183–214 (2013).
9Connor, T., and Wong, H., “Emergent Properties,” The Stanford
Encyclopedia of Philosophy, Spring 2012 Ed., E. N. Zalta (ed.), last
modified February 12, 2012, http://plato.stanford.edu/entries/
properties-emergent/.
10Alberts, D. S., and Hayes, R. E., Understanding Command and Control,
Command and Control Research Program Publication Series, Washington, DC (2006).
11U.S. Navy Information Dominance Roadmap, 2013–2028, U.S. Navy,
Washington, DC (Mar 2013).
12Seymour, J. S., and O’Driscoll M. J., “APL Applied Systems Engineering: Guest Editors’ Introduction,” Johns Hopkins APL Tech. Dig. 29(4),
306–309 (2011).
13Valinoto, T., “Application of MBSE Theory in a World of Practical
Deadlines and Deliverables: Lessons Learned,” in Proc. MBSE Symp.:
INCOSE Chesapeake Chapter and JHU/APL (29 Mar 2014).
14Defense Advanced Research Projects Agency (DARPA), DARPABAA-14-02, “Communications in Contested Environments (C2E),”
Solicitation Number DARPA-BAA-14-02 (Jan 2014).
15Open
Group, Technical Standard for Future Airborne Capability Environment (FACETM), Edition 2.0 (Feb 2013), http://www.opengroup.
org/face/tech-standard-2.
16U.S. Department of Defense, “Waveform Specification JTRS Software Waveform,” Tactical Targeting Networking Technology (TTNT),
Draft Revision 4.0 (Jul 2005).
17Object Management Group, “Documents Associated With SysML
Version 1.2,” http://www.omg.org/spec/SysML/1.2/.
18Object Management Group, “Index of /spec/UML/2.3/Superstructure,”
http://www.omg.org/spec/UML/2.3/Superstructure/.
19Friedenthal, S., Griego, R., and Sampson, M., “INCOSE Model
Based Systems Engineering (MBSE) Initiative,” in Proc. INCOSE
2007 Symp., San Diego, CA, p. 17 (2007), https://www.incose.org/
enchantment/docs/07Docs/07Jul_4MBSEroadmap.pdf.
20Hause, M., Stuart, A., Richards, D., and Holt, J. “Testing Safety Critical Systems with SysML/UML,” in Proc. 15th IEEE Intl. Conf. on Engineering of Complex Computer Systems, Oxford, UK, pp. 325–330 (2010).
21National Defense Industrial Association (NDIA) Systems Engineering Division M&S Committee, “Final Report of the Model
Based Engineering (MBE) Subcommittee,” NDIA, Arlington, VA
(10 Feb 2011).
THE AUTHORS
K. Dewayne Brown is a communications systems engineer and Supervisor of the Resilient Tactical Communications
Section in APL’s Asymmetric Operations Sector. He provided MBSE expertise for the ADI C2E and netted ISR projects.
Mary Beth A. Chipkevich is a Principal Professional Staff member in the Force Projection Sector and Program Manager
for Unmanned Systems and Autonomous Technologies in the Precision Strike Mission Area. Robert J. Bamberger is a
member of APL’s Principal Professional Staff and Supervisor of the Autonomy Section in the Research and Exploratory
Development Department. He was the Technical Lead for the Defense Advanced Research Projects Agency’s Airborne
Dominance Initiative (ADI) Communications in a Contested Environment (C2E) project. Tam-Thanh C. Huang is a
member of APL’s Senior Professional Staff and is a communications engineer with extensive experience in softwaredefined radio design and implementation of various military waveforms. Mark A. Matties is a member of APL’s Senior
Professional Staff in the Communication and Networking Systems Group and Supervisor of the Advanced Software
Defined Networking Section. He currently researches the security and advanced applications of software-defined
networking to provide secure, resilient networks. James D. Reeves is a member of APL’s Senior Professional Staff and
is a systems, network, and information systems security engineer for advanced technologies in the Aerospace Systems
Analysis Group of the Precision Strike Mission Area. He was the Lead Systems Engineer for the netted ISR project.
Christopher A. Rouff is a member of APL’s Senior Professional Staff in the Communication and Networking Systems
Group. He is a computer scientist doing research in automatic code generation from system models to quickly produce
simulations and prototypes of new system concepts. The ADI and netted ISR teams can be contacted through the
Program Manager, Mary Beth Chipkevich. Her e-mail address is mary.beth.chipkevich@jhuapl.edu.
36­­­­
Johns Hopkins APL Technical Digest, Volume 33, Number 1 (2015), www.jhuapl.edu/techdigest
Download